#infoq — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #infoq, aggregated by home.social.
-
“Just add more instructions to the prompt.” Sound familiar?
When an LLM acts up, prompt tuning can feel like the obvious fix. But context engineering isn’t about telling the model more - it’s about controlling what it actually sees.
👀 Sneak peek: Ricardo Ferreira, Principal Developer Advocate at Redis, shares hard-learned lessons from the frontlines of AI engineering.
🗓️ Interested to learn more? Mark your calendar for September 02, 2026 as this video will be live on InfoQ!
-
5,500 access requests daily. 10,000 employees.
To scale Just-In-Time Access (JITA), HubSpot redesigned its authorization system with a rule engine architecture.
Access requests are evaluated through independent rules organized as a directed acyclic graph (DAG), adding structured decision metadata, rule-level observability, and governance workflows.
Learn how the architecture improves authorization at scale: https://bit.ly/4fS7EbI
#SoftwareArchitecture #Authorization #Observability #Security #InfoQ
-
Are modern #LLMs ready for root cause analysis?
A new Coroot experiment across 11 models suggests yes - provided they get the right context.
Key takeaway: The hard problem is shifting away from model capability and onto the pipelines that correlate telemetry.
🔗 Dive deeper: https://bit.ly/4fBhL4r
-
Are modern #LLMs ready for root cause analysis?
A new Coroot experiment across 11 models suggests yes - provided they get the right context.
Key takeaway: The hard problem is shifting away from model capability and onto the pipelines that correlate telemetry.
🔗 Dive deeper: https://bit.ly/4fBhL4r
-
Are modern #LLMs ready for root cause analysis?
A new Coroot experiment across 11 models suggests yes - provided they get the right context.
Key takeaway: The hard problem is shifting away from model capability and onto the pipelines that correlate telemetry.
🔗 Dive deeper: https://bit.ly/4fBhL4r
-
Are modern #LLMs ready for root cause analysis?
A new Coroot experiment across 11 models suggests yes - provided they get the right context.
Key takeaway: The hard problem is shifting away from model capability and onto the pipelines that correlate telemetry.
🔗 Dive deeper: https://bit.ly/4fBhL4r
-
Are modern #LLMs ready for root cause analysis?
A new Coroot experiment across 11 models suggests yes - provided they get the right context.
Key takeaway: The hard problem is shifting away from model capability and onto the pipelines that correlate telemetry.
🔗 Dive deeper: https://bit.ly/4fBhL4r
-
Login-time #Authorization leaves sensitive cloud data exposed mid-session.
In this #InfoQ article, Venkata Nedunoori examines a Continuous Authorization Architecture built around:
• Risk-based policy evaluation
• Behavioral baselines
• Privacy-preserving audit trails
• A phased implementation strategy🔗 Read now: https://bit.ly/44rmyjL
-
💡 72% of organizations have adopted AI, yet most are stuck in “pilot purgatory”!
The reason? It’s rarely a technology problem- it’s a flow problem.
As this #InfoQ article explores, AI is a force multiplier: if your organization is dysfunctional, AI will simply amplify that dysfunction at scale. 📈
The real shift isn’t just #AIadoption - it’s moving toward #FastFlow by: ✅ Defining clear domain boundaries ✅ Engineering context
📖 Read the deep dive: https://bit.ly/4qdZMVt
#SoftwareArchitecture #AIArchitecture #AI
*******************************************
𝑇ℎ𝑖𝑠 𝑎𝑟𝑡𝑖𝑐𝑙𝑒 𝑤𝑎𝑠 𝑤𝑟𝑖𝑡𝑡𝑒𝑛 𝑏𝑦 𝑝𝑎𝑟𝑡𝑖𝑐𝑖𝑝𝑎𝑛𝑡𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑜𝑛𝑙𝑖𝑛𝑒 𝐼𝑛𝑓𝑜𝑄 𝐶𝑒𝑟𝑡𝑖𝑓𝑖𝑒𝑑 𝐴𝑟𝑐ℎ𝑖𝑡𝑒𝑐𝑡 𝑃𝑟𝑜𝑔𝑟𝑎𝑚. 𝐼𝑡 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑠 𝑡ℎ𝑒 𝑐𝑎𝑝𝑠𝑡𝑜𝑛𝑒 𝑜𝑓 𝑡ℎ𝑒𝑖𝑟 𝑤𝑜𝑟𝑘, 𝑟𝑒𝑓𝑙𝑒𝑐𝑡𝑖𝑛𝑔 𝑡ℎ𝑒 𝑐𝑜ℎ𝑜𝑟𝑡'𝑠 𝑐𝑜𝑙𝑙𝑒𝑐𝑡𝑖𝑣𝑒 𝑙𝑒𝑎𝑟𝑛𝑖𝑛𝑔𝑠 𝑜𝑛 𝑡ℎ𝑒 𝑖𝑛𝑡𝑒𝑟𝑠𝑒𝑐𝑡𝑖𝑜𝑛 𝑜𝑓 𝐴𝐼 𝑎𝑛𝑑 𝑚𝑜𝑑𝑒𝑟𝑛 𝑠𝑜𝑓𝑡𝑤𝑎𝑟𝑒 𝑎𝑟𝑐ℎ𝑖𝑡𝑒𝑐𝑡𝑢𝑟𝑒. -
🚀 #MSTest 3.4 is out now!
Key updates include improvements to MSTest.Analyzers, MSTest.Sdk, and MSTest.Runner. Plus, added support for WinUI applications!
Find out all the details on #InfoQ 👉 https://bit.ly/3yWE9DH
-
The Abu Dhabi government's Technology Innovation Institute released #Falcon180B - the largest openly available #LargeLanguageModel.
It contains 180 billion parameters and outperforms GPT-3.5 on the MMLU benchmark.
More on #InfoQ: https://bit.ly/3RjydM7
-
The Abu Dhabi government's Technology Innovation Institute released #Falcon180B - the largest openly available #LargeLanguageModel.
It contains 180 billion parameters and outperforms GPT-3.5 on the MMLU benchmark.
More on #InfoQ: https://bit.ly/3RjydM7
-
The Abu Dhabi government's Technology Innovation Institute released #Falcon180B - the largest openly available #LargeLanguageModel.
It contains 180 billion parameters and outperforms GPT-3.5 on the MMLU benchmark.
More on #InfoQ: https://bit.ly/3RjydM7
-
The Abu Dhabi government's Technology Innovation Institute released #Falcon180B - the largest openly available #LargeLanguageModel.
It contains 180 billion parameters and outperforms GPT-3.5 on the MMLU benchmark.
More on #InfoQ: https://bit.ly/3RjydM7
-
The Abu Dhabi government's Technology Innovation Institute released #Falcon180B - the largest openly available #LargeLanguageModel.
It contains 180 billion parameters and outperforms GPT-3.5 on the MMLU benchmark.
More on #InfoQ: https://bit.ly/3RjydM7
-
A question every #mobileapp project must ponder is:
Do we want to use #NativeApp, or do we want to use a #CrossPlatform framework? And if you go the cross-platform route, which framework better suits our needs?
➡️ Hear about frameworks for mobile development, pros & cons
➡️ Learn how to make the best decision for a mobile app-building framework🎥 #InfoQ video: http://bit.ly/3ZFCFWG